Passive optical networks (PON) have gained much interest in the past years because of their fiber gain, broadband capability, and unpowered outside plant which reduces maintenance cost [1]. (Note that in the above and following description, a reference's identification [e.g., 1] refers to that reference's location in the Appendix) Generally, PONs can be either time division multiplex (TDM), wavelength division multiplex (WDM) or both. In a TDM PON, the signals, which are electronically multiplexed at the Central Office (CO), are equally split at the remote node (RN) by a passive power splitter among all the optical network units (ONU) where they are electronically demultiplexed. The receivers at each ONU have to process the information at the aggregate bit-rate. Furthermore, the optical signal power coming from the CO is attenuated at the RN because of the splitting loss. In WDM PONs, each ONU is assigned a different wavelength, which is passively demultiplexed at the RN by a router. This provides a virtual point-to-point connectivity which also means that the receiver and the transmitter do not operate at the aggregate bit-rate. Such a system allows high flexibility in bandwidth allocation and upgradability for individual ONUs or for the whole system. However, one of the prices to pay for this flexibility is wavelength management. Each wavelength must be precisely spaced and aligned with the router. The multifrequency laser (MFL) provides a precise channel spacing due to its internal waveguide grating router (WGR) which acts as an intra cavity filter [2]. The WGR also provides single-knob wavelength comb tunability by changing the device temperature alone, simplifying the locking of the MFL source wavelengths to the passive WGR demultiplexer at the RN, which drifts due to temperature changes of the uncontrolled outside plant. What is needed is a practical and reliable way to lock the MFL wavelength frequencies to the WGR demultiplexer with temperature changes in the outside plant (i.e., WGR).